1. Field of the Invention
Generally, the present invention relates to the field of fabricating microstructures, such as integrated circuits, micromechanical structures and the like, and, more particularly, to the formation of an ultra-thin dielectric oxide layer having increased resistance against migration of charge carriers through the dielectric oxide layer.
2. Description of the Related Art
Presently, microstructures are integrated into a wide variety of products. One example in this respect is the employment of integrated circuits that, due to their relatively low cost and high performance, are increasingly used in many types of devices, thereby allowing superior control and operation of those devices. Due to economic reasons, manufacturers of microstructures, such as integrated circuits, are confronted with the task of steadily improving performance of these microstructures with every new generation appearing on the market. However, these economic constraints not only require improving the device performance but also demand a reduction in size so as to provide more functionality of the integrated circuit per unit chip area. Thus, in the semiconductor industry, ongoing efforts are being made to reduce the feature sizes of feature elements. In present-day technologies, the critical dimension of these elements approach 0.1 xcexcm and less. In producing circuit elements of this order of magnitude, along with many other issues especially arising from the reduction of feature sizes, process engineers face several issues. For example, one such issue involves providing extremely thin dielectric layers on an underlying material layer, wherein certain characteristics of the dielectric layer, such as permittivity and/or resistance against charge carrier tunneling and the like, have to be improved without sacrificing the physical properties of the underlying material layer.
One important example in this respect is the formation of ultra-thin gate insulation layers of field effect transistors, such as MOS transistors. The gate dielectric of a transistor has an essential impact on the performance of the transistor. As is commonly known, reducing the size of a field effect transistor, that is reducing the length of a conductive channel that forms in a portion of a semiconductor region by applying a control voltage to a gate electrode formed on the gate insulation layer, also requires the reduction of the thickness of the gate insulation layer to maintain the required capacitive coupling from the gate electrode to the channel region. Currently, most of the highly sophisticated integrated circuits, such as CPUs, memory chips and the like, are based on silicon and, therefore, silicon dioxide has preferably been used as the material for the gate insulation layer due to the well-known and superior characteristics of the silicon dioxide/silicon interface. For a channel length on the order of 100 nm and less, however, the thickness of the gate insulation layer has to be reduced to about 2 nm in order to maintain the required controllability of the transistor operation. Steadily decreasing the thickness of the silicon dioxide gate insulation layer, however, leads to an increased leakage current therethrough, thereby resulting in an unacceptable increase of static power consumption as the leakage current exponentially increases for a linear reduction of the layer thickness.
Therefore, great efforts are presently being made to replace silicon dioxide by a dielectric exhibiting a higher permittivity so that a thickness thereof may be higher than the thickness of a corresponding silicon dioxide layer providing the same capacitive coupling. A thickness for obtaining a specified capacitive coupling will also be referred to as capacitive equivalent thickness and determines the thickness that would be required for a silicon dioxide layer. It turns out, however, that it is difficult to incorporate high-k materials into the conventional integration process and, more importantly, the provision of a high-k material as a gate insulation layer seems to have a significant influence on the carrier mobility in the underlying channel region, thereby remarkably reducing the carrier mobility and thus the drive current capability. Hence, although an improvement of the static transistor characteristics may be obtained by providing a thick high-k material, at the same time, an unacceptable degradation of the dynamic behavior presently makes this approach less than desirable.
A different approach that is currently favored is the employment of an integrated silicon oxide/nitride layer stack that may reduce the gate leakage current by 0.5 to 2 orders of magnitude while maintaining compatibility with standard CMOS process techniques. It has been found that the reduction of the gate leakage current mainly depends upon the nitrogen concentration incorporated into the silicon dioxide layer by means of plasma nitridation. Although this approach seems to relax the issue of gate dielectric leakage for the present circuit generation, this approach does not seem to allow further aggressive dielectric thickness scaling required for future device generations. To more clearly demonstrate the problems involved in the conventional process technique, a typical process flow for forming a gate insulation layer, including a nitride/silicon dioxide layer, will now be described with reference to FIGS. 1a-1e. 
In FIG. 1a, a semiconductor device 100 comprises a silicon substrate 101, in which an active region 103 is defined by shallow trench isolations 102. A thin dielectric base layer 110, for example formed of a grown oxide layer, covers the active region 103. Moreover, the semiconductor device 100 is exposed to a nitrogen-containing plasma, indicated by reference sign 104.
Typically, the semiconductor device 100 may be formed according to the following process sequence. After formation of the shallow trench isolation 102 and various implantation steps to generate a required well dopant profile (not shown) in the active region 103, the dielectric base layer 110 is formed by a conventional oxidation process or by a rapid thermal oxidation process. Subsequently, the semiconductor device 100 is exposed to the nitrogen-containing plasma 104 to introduce nitrogen ions into the silicon dioxide layer 110 to improve, as explained above, the resistance of the dielectric base layer 110 against charge carrier migration and to increase the permittivity thereof. An energy of the ions in the nitrogen-containing plasma 104 is substantially determined by the difference between the plasma potential and the floating potential of the semiconductor device 100, wherein this voltage is difficult to adjust or even may not be adjustable at all. The ion density and, thus, the nitridation rate depends on process parameters, such as the high frequency (HF) power, the temperature of the plasma ambient 104, the pressure thereof, and the like.
As is well known, nitrogen atoms, introduced into the active region 103 and thus into the channel region of the transistor device to be formed, significantly affect the electrical characteristics of the transistor device in that both the crystallinity of the active region 103 is deteriorated and the charge carrier mobility is degraded. Consequently, the introduction of nitrogen into the active region 103 has to be suppressed as much as possible in view of a required high transistor performance. On the other hand, a thickness of the dielectric base layer 110 is to be scaled down in conformity with the device dimensions which would, however, at a certain minimum dielectric thickness, lead to an increased penetration of nitrogen ions into the active region 103 during the plasma treating 104. As a consequence, there exists a severe trade-off between the improvement of the transistor device performance by scaling down the dielectric base layer 110 and the device degradation caused by the incorporation of nitrogen into the active region 103. FIGS. 1b and 1c will more clearly illustrate this situation.
In FIG. 1b, on the left-hand side, a cut-out of FIG. 1a is schematically illustrated in a magnified view, in which the dielectric base layer 110 includes nitrogen atoms according to a concentration profile 112 along a depth direction 111. As is evident from FIG. 1b, a thickness 113 of the dielectric base layer 110 is selected so that a penetration of nitrogen into the underlying active region 103 is minimal. The right-hand side of FIG. 1b depicts a graph, wherein the concentration profile 112 is plotted versus the depth direction 111. As can be seen from this graph, the nitrogen concentration drops to a very low value, which in the present illustration is idealized as 0, within the thickness 113 of the base layer 110, thereby substantially avoiding any device degradation caused by the reduction of carrier mobility. FIG. 1b represents the situation when the design thickness 113 is in conformity with a desired channel length and suffices to still allow a substantial blocking of the nitrogen.
FIG. 1c, on the other hand, illustrates the situation when a thickness 113xe2x80x2 of the base layer 110 has to be reduced in conformity with design rules compared to that of FIG. 1b and, therefore, the concentration profile 112 reaches down into the active area 103. The right-hand side of FIG. 1c shows a resulting nitrogen concentration 112 with respect to the depth direction 111 and clearly reveals that a significant amount of nitrogen is in the active region 103, to thereby negatively affect the carrier mobility.
FIG. 1d schematically shows the semiconductor device 100 in an advanced manufacturing stage. Drain and source regions 107 are formed within the active region 103 and a gate electrode 106 is formed on the patterned dielectric base layer 110, which is now indicated as 110a, wherein the gate insulation layer 110a has the thickness 113 and a nitrogen concentration profile 112 as shown in FIG. 1b. Moreover, sidewall spacers 105 are formed adjacent to the gate electrode 106.
Typical process steps for forming the semiconductor device 100, as shown in FIG. 1d, include well-known advanced photolithography and etch techniques, as well as implantation steps, and thus a detailed description thereof will be omitted.
FIG. 1e, on the other hand, schematically depicts the semiconductor device 100 having the gate insulation layer 110a with the reduced thickness 113xe2x80x2 as shown in FIG. 1c, so that a corresponding residual nitrogen concentration is present within the drain and source regions 107 and the relevant portion of the active region 103.
As a consequence, the prior art processing described above allows a scaling of the thickness 113 of the gate insulation layer 110a to a value that substantially prevents penetration of nitrogen into the active region 103, thereby obtaining improved device performance. However, when design requirements demand a further scaling of the thickness 113 to conform to the corresponding transistor dimensions, i.e., when the design capacitive equivalent thickness requires the thickness 113xe2x80x2, an unacceptable amount of nitrogen is introduced into the surface portion of the active region 103 so that the reduced carrier mobility may result in a device performance degradation.
In view of the problems explained above, it is thus highly desirable to provide a technique allowing for improvement of the resistance of an oxide/nitride layer stack against migration of charge carriers therethrough without unduly negatively affecting the physical characteristics of an underlying material layer, such as the carrier mobility of a silicon layer.
The present invention is based on the finding that the nitrogen concentration introduced into a material layer underlying an oxide layer that is to be doped to improve with respect to permittivity and resistance against a charge carrier tunneling may effectively be reduced in that the nitrogen ion concentration during a plasma nitridation is controlled such that at an initial phase of the process a low ion density is established. During the further progression of the process, the thickness of the oxide layer may increase due to the further oxidation and/or deposition of residual oxygen and the nitrogen ion density is then increased so as to incorporate a required amount of nitrogen into the oxide layer. Especially the pressure during the plasma nitridation may be selected for controlling the nitrogen ion density in the plasma ambient.
According to one illustrative embodiment of the present invention, a method of forming an insulation layer includes the formation of an oxide base layer on a substrate. The substrate is then introduced into a nitrogen-containing plasma ambient of controlled pressure, wherein the pressure is reduced during exposure of the substrate to the nitrogen-containing plasma ambient.
According to another illustrative embodiment of the present invention, a method of forming an insulation layer comprises forming an oxide layer with a first thickness on a semiconductor region and oxidizing the semiconductor region in a nitrogen-containing plasma ambient by oxygen residuals that are present in the nitrogen-containing plasma ambient. Moreover, a nitrogen ion density is increased while the semiconductor region is oxidized.
According to still a further illustrative embodiment of the present invention, a method of forming a gate insulation layer comprises forming a base oxide layer on an active region with a thickness that is less than a predefined design thickness of the gate insulation layer. The substrate is then introduced in a nitrogen-containing plasma ambient of controlled pressure, wherein the pressure is reduced during the exposure of the substrate to the plasma ambient while simultaneously the thickness of the base oxide layer is increased by residual oxygen in the plasma ambient.
Pursuant to yet another illustrative embodiment of the present invention, a method of controlling a nitrogen concentration during the formation of an insulation layer is provided. The method includes the exposure of a substrate to a nitrogen-containing plasma ambient, wherein the substrate has formed thereon an oxide layer of a predefined thickness. The thickness of the oxide layer is then monitored and a monitor signal is generated that indicates the monitored thickness. Additionally, at least one process parameter determining a nitrogen ion density in the plasma ambient is controlled on the basis of the monitor signal.
In a further illustrative embodiment of the present invention, a system for forming an insulation layer comprises a process chamber and a plasma generator arranged therein. Moreover, a controllable vacuum generator is provided and is configured to adjust a pressure in the process chamber. Moreover, a thickness measurement tool is provided that is configured to determine the thickness of a layer formed above a substrate positioned within the process chamber. Finally, a control unit is connected to the controllable vacuum generator and the thickness measurement tool and is, furthermore, adapted to receive a monitor signal from the thickness measurement tool and to control the vacuum generator on the basis of the received monitor signal.